Testing galaxy evolution theories using galaxy gradients and machine learning

Galaxy spectra are a rich source of information about the properties of gas and stars in the galaxies. The development of large integral field spectroscopic surveys has allowed for the spatially resolved observations of thousands of galaxies in the local Universe. The advantage of such surveys lies on the ability to compute the radial profiles of different galaxy properties, which plays a key role in understanding how galaxies evolve. On the other hand, large single fibre spectroscopic surveys, such as the Dark Energy Spectroscopic Instrument (DESI) survey, play a crucial role in understanding the underlying structure of the Universe. To fully explore galaxy evolution science with such surveys, fibre aperture corrections need to be computed to account for the fact that the fibre only covers a fraction of the total area of each galaxy. This thesis makes use of spatially resolved data to investigate the radial profiles of galaxies, not only to constrain theories of galaxy evolution but also to compute fibre aperture corrections for single fibre surveys. The shape and gradient of these profiles are investigated to test how both internal and external processes shape the evolution of galaxies. We use the integral field unit (IFU) spectroscopic data from the Mapping Nearby Galaxies at APO (MaNGA) survey to compute the point spread function-corrected gradients of different galaxy properties, such as the stellar mass surface density, mass-weighted age and metallicity and specific star formation rate. Furthermore, we use the random forest permutation importance to understand which galaxy properties play a more significant role in setting the observed gradients and discuss its implications to models of galaxy evolution (Chapter 3). With the results obtained from machine learning, we present a new method for computing aperture corrections to the four different properties explored in this work, which is based on splitting galaxies into different types given the features highlighted by the random forest algorithm (Chapter 4). Finally, we investigate how galaxy properties are affected by their large scale environment, the cosmic web, with a specific focus on how the alignment between galaxies and their host filament correlates with different galaxy properties (Chapter 5).